Analysis of acute pancreatitis associated with SGLT-2 inhibitors and predictive factors of the death risk: Based on food and drug administration adverse event report system database
Background and objectives: The US FDA and Health Canada have successively published potential red flags for acute pancreatitis caused by sodium-dependent glucose transporter 2 inhibitors (SGLT-2i). However, existing studies have focused on case reports. We aimed to assess the possible association of SGLT-2i with acute pancreatitis by analyzing postmarketing adverse events reported in the FDA adverse event reporting system (FAERS), to explore risk factors for SGLT-2i-related acute pancreatitis death, and to build a nomogram. Methods and Results: We used a disproportionality analysis of suspected acute pancreatitis-related reports in the FAERS database of patients from the use of SGLT-2i from the first quarter of 2013 to the fourth quarter of 2021. Single-factor and multi-factor logistic regression analyses were performed using the relevant clinical information of patients, and risk factors were combined with the age of patients to construct a SGLT-2i risk prediction model for acute pancreatitis-related death. A total of 757 reports were retrieved. The largest number of acute pancreatitis-related cases were caused by canagliflozin (317 reports), which was also the strongest agent associated with acute pancreatitis, with the information component (IC 2.41, lower 95% one-sided confidence interval 2.16), the reporting odds ratio (ROR 5.37, 95% two-sided confidence interval 4.8–5.99), and the empirical Bayesian geometric mean (EBGM 5.32, lower 90% one-sided confidence interval 4.85). The median time to acute pancreatitis was 54 (interquartile range [IQR] 14–131) days, and approximately 83% of adverse events occurred within 6 months. Odds ratio(OR) adjusted by acute pancreatitis and the coadministration of SGLT-2i with dipeptidyl peptidase 4 inhibitor (DPP-4i), glucagon-like peptide 1 analog (GLP-1RA), and angiotensin converting enzyme inhibitor (ACEIs) was 1.39, 1.97, and 1.34, respectively, all of which were statistically significant. Logistic regression analysis showed that different SGLT-2i type and their combinations with statins were independent risk factors for acute pancreatitis mortality in the patients ( p < 0.05). The mortality risk prediction model showed good discrimination and clinical applicability in both the training set (AUC 0.708) and the validation set (AUC 0.732). Conclusion: SGLT-2i may increase the risk of acute pancreatitis especially within the first 6 months of drug administration. Combination with DPP-4i, GLP-1RA or ACEIs significantly increases the risk of acute pancreatitis. In addition, different SGLT-2i type and their combination with statins are risk factors that can predict the risk of death following acute pancreatitis.